Peer loss timeout

Sander Kuusemets <sander.kuusemets@xxxxx> · Tue, 16 Feb 2016 13:10:06 +0200



    Hello,

    
    I have a problem with a two node setup, that are replica 2 servers,
    and also clients.

    
    Actions on node1:

    
      gluster peer probe node2

      gluster volume create fs replica 2 transport tcp node1:/fs
      node2:/fs force

      gluster volume start fs

      mount -t glusterfs node1:/fs /test
    

    Actions on node2:

    
    gluster peer probe node1

      mount -t glusterfs node2:/fs /test
    

    Everything works fine, except when one of the peers are shut down,
    then the other peer basically freezes for about a minute. Any action
    waits some kind of a timeout, even time ls / gives me
    an output of ~1m10s.

    
    I tried gluster volume set fs network.ping-timeout "5" but
    this did not change a thing.

    
    I'd like this delay to be less than 10 seconds. Is this possible?

    
    Best regards,

    -- 
Sander Kuusemets
University of Tartu, High Performance Computing, IT Specialist


_______________________________________________
Gluster-users mailing list
Gluster-users@xxxxxxxxxxx
http://www.gluster.org/mailman/listinfo/gluster-users